CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   Main CFD Forum (http://www.cfd-online.com/Forums/main/)
-   -   Whose programming is more difficult?DNS or LES? (http://www.cfd-online.com/Forums/main/10581-whose-programming-more-difficult-dns-les.html)

Adonis January 6, 2006 10:09

Whose programming is more difficult?DNS or LES?
 
I want to choose one requiring less effort.

Guillaume January 6, 2006 10:11

Re: Whose programming is more difficult?DNS or LES
 
DNS, by definition.

tom January 6, 2006 10:40

Re: Whose programming is more difficult?DNS or LES
 
DNS is computationally expensive and therefore requires perhaps parallelisation and the handling of much data.LES is cheaper and probably can be carried out on a single fast PC; no parallelisation is then required.

Thus, developing and running a DNS code does not have to be easier than developing and running an LES code ...

Renato. January 6, 2006 11:20

Re: Whose programming is more difficult?DNS or LES
 
DNS, programming cheap (by definition as Guillaume said) but computationally very expensive

LES with classic Smagorisky, programming and computationally cheap (but with little accuracy)

Regards

Renato.

Renato. January 6, 2006 11:24

Rephrasing
 
DNS: programming cheap (by definition as Guillaume said) but computationally very expensive

LES + classic Smagorisky and without any wall law: programming and computationally cheap (but with little accuracy)

Regards

Renato.


Guillaume January 6, 2006 17:23

Re: Whose programming is more difficult?DNS or LES
 
I understand the arguments about intensive computation for DNS with respect to LES but, as one does not use LES and DNS for the same cases (dimensions, turbulence intensity), in practice, one barely manages to resolve the inertial range of the turbulence spectrum without a fine mesh. Thus, LES is expensive as well.



All times are GMT -4. The time now is 10:35.